Scores predictions for a trained classification or regression model
Category: Machine Learning / Score
Applies to: Machine Learning Studio
This content pertains only to Studio. Similar drag and drop modules have been added to the visual interface in Machine Learning service. Learn more in this article comparing the two versions.
This article describes how to use the Score Model module in Azure Machine Learning Studio, to generate predictions using a trained classification or regression model.
How to use Score Model
Add the Score Model module to your experiment in Studio.
Attach a trained model and a dataset containing new input data.
The data should be in a format compatible with the type of trained model you are using. The schema of the input dataset should also generally match the schema of the data used to train the model.
Run the experiment.
After you have generated a set of scores using Score Model:
- To generate a set of metrics used for evaluating the model’s accuracy (performance). you can connect the scored dataset to Evaluate Model,
- Right-click the module and select Visualize to see the a sample of the results.
- Save the results to a dataset.
The score, or predicted value, can be in many different formats, depending on the model and your input data:
- For classification models, Score Model outputs a predicted value for the class, as well as the probability of the predicted value.
- For regression models, Score Model generates just the predicted numeric value.
- For image classification models, the score might be the class of object in the image, or a Boolean indicating whether a particular feature was found.
Publish scores as a web service
A common use of scoring is to return the output as part of a predictive web service. For more information, see this tutorial on how to create a web service based on an experiment in Azure ML Studio:
- Compare Binary Classification Models
- Compare Multiclass Classification Models
- Compare Multiple Regression Models
Models not supported by Score Model
If you are using one of the following special types of model, you might need to use one of these custom scoring modules:
Create recommendations or generate data for evaluating a recommender: Use Score Matchbox Recommender
If the data that you are scoring contains missing values, in many cases no score will be generated for the entire row.
The following machine learning models require that data have no missing values. When using the following machine learning models, review the data before passing it to Score Model, and use Clean Missing Data to amend the missing values in input columns.
|Trained model||ILearner interface||Trained predictive model|
|Dataset||Data Table||Input test dataset|
|Scored dataset||Data Table||Dataset with obtained scores|
|Error 0032||Exception occurs if argument is not a number.|
|Error 0033||Exception occurs if argument is Infinity.|
|Error 0003||Exception occurs if one or more of inputs are null or empty.|
|Error 0013||Exception occurs if the learner that is passed to the module is an invalid type.|